Analytical performance goals for measuring prostate-specific antigen.

We have assessed the feasibility of using fixed-limit criteria based on medical relevance and biological variation for evaluating the analytical performance of the prostate-specific antigen (PSA) test. The estimated within-subject variation of serum PSA is on the order of 10-20% at clinical decision points. The calculated performance goals of 5-10% CV are attainable with current immunoassay technology and agree with precision goals based on clinical experience and the current clinical use of the test. However, new clinical applications of PSA may require a degree of analytical performance that current methods may not be able to provide. The PSA model demonstrates the need for biologically based fixed-limit criteria for all tumor-marker tests.